Blog

Found some old video

posted Jun 24, 2014, 3:24 AM by Evan Morrison

Just found some old videos that I made last year being tweeted about. These are on how to setup and deploy the Activiti workflow engine on your home pc and then deploy a process that includes drools rules.

I'm planning on doing a re-recording of these soon, with the latest version. I'm very excited to hear that Activiti now has an enterprise release .


YouTube Video


YouTube Video



A different take on process similarity

posted May 28, 2014, 3:13 PM by Evan Morrison

On Monday 26th of May, I gave a talk at Cafe DSL. Cafe DSL is a weekly research seminar series run by the Decision System Lab at the University of Wollongong.

In this talk I presented a new take on measuring process similarity using natural language and a novel combination of distance measures with a refined instance of bisimulation. You can download my slides for the talk at:

Book Review : Activiti 5.x Business Process Management Beginners Guide

posted Apr 20, 2014, 6:17 PM by Evan Morrison   [ updated Apr 22, 2014, 2:44 AM ]

Laliwala, Z, Mansuri, I (2014) Activiti 5.x Business ProcessManagement Beginners Guide, Packt Publishing

Dr. Zakir Laliwala and Irshad Mansuri have written an exceptional introduction to the Activiti business process engine in “Activiti 5.x Business Process Management”. The book is a fantastic companion to the Activiti user manual and Activiti in Action. The book is written with the beginner process analyst in mind. It has lots of examples showing hands-on methods for using Activiti.  The book is extremely accessible and little theoretical or programming knowledge is assumed.

The book covers all topics needed to get started with Activiti. It first demonstrates installation on a user’s machine and then all the way to implementing advanced workflows and complex event processing. This book is best read front to back, and is not intended as a reference book.

One critic of the book is that relies on heavy use of screenshots. These are great for helping a reader test an example. Though it means that the reader is sometimes left puzzled by what each step is  doing. For example, when explaining how to deploy a process, we  are told to generate a .bar. It then shows how to make the file using the eclipse plug-in; but, there is no description of what a .bar file is or why it is needed. A further reading sections in each chapter would be an improvement.

Overall the book will guide a beginning Activiti user to a point where they understand basic terminology will be able to create and deploy a process on Activiti. The style of writing used for this book does make getting started with Activiti very easy.


You can find the book at http://www.packtpub.com/activiti-bpm-beginners-guide/book

Literate Java Posets

posted Mar 25, 2014, 10:17 PM by Evan Morrison

I've loved the idea and concept of literate programming since I first came across it in my undergraduate years. The problem I've always had with it is that I've never had a practical use for it. Until now that is. Through the write up of my thesis I've started to document the decision procedures for some of the more complex functions in TextSeer, including accumulation as a belief update function, abduction, and default reasoning. 

Those functions will all be posted here sooner or later, though in the mean time to practice my literate coding I took an opportunity to write up a poset class I had to write today. 

I'm currently implementing a few instances of the c-semiring framework, to do that I've had to use partially ordered sets (poset). After a brief search for a java poset class, I couldn't find a working copy online so I wrote one. I've attached a copy of the literate description of the poset class which is now in TextSeer (https://github.com/edm92/TextSeer/blob/master/src/PosetExample.java) 


Deductive Closure

posted Mar 10, 2014, 7:17 PM by Evan Morrison

Find out about my latest development on TextSeer at http://www.evanmorrison.com.au/1/post/2014/03/deductive-closure.html. As you will see I've been developing a deductive closure operation to make the default logic reasoner give more accurate results.

Lealana Crypto Coins are In!

posted Jan 31, 2014, 1:15 PM by Evan Morrison   [ updated Feb 16, 2014, 2:46 PM ]

About a year ago I posted about Bitcoins and the future of money on the web. At the time Bitcoin was a month away from the first boom where it sky rocketed to the dizzying heights of $300USD. Of course since then it has been a bumpy ride. The currency has gone up and done like a yoyo. In June it was back down at $75 and then by September it had breached the $1000USD mark. Now personally I’m extremely happy that Bitcoin continues to rise; however, with high volatility in the market, it is extremely hard to convince shop owners to accept Bitcoin. As such I really believe that there is a need for at least one other crypto coin to cushion the volatility (which a number of ‘investors’ are actually taking advantage of), and in this case that coin is Litecoin. Litecoin is a Bitcoin variant developed by Charles Lee that is similar to the original with the added properties of faster transaction processing, and the ability to produce four times as many coins (to ensure that the currency holds value, there is a limit on the number of coins that can be generated). As a person interested in the propagation of crypto currencies, I converted a handful of my original Bitcoins into Litecoin which I send out as donations and use to access various VPN services.

Back before the big run on Bitcoin I stumbled across a website selling minted Bitcoins, the Casascius coin made in the US by Mike Caldwell. Casascius coins are physically minted coins that have attached to them a holographic sticker that contains an address to a physical Bitcoin, which can be redeemed at some point in the future. At the time I was interested in purchasing the coins as collectables though was put off by the high pricing which is laughable in hindsight as the physical coins now sell for ten times the value of the digital coin as collectors’ items. What has made this worse is the shutdown of Mike’s business under AML concerns. So when in December I saw the news reports of Mike’s troubles I started to hunt for Litecoin minters to start my own collection before it is too late (again).

After a quick search I found a cool Hawaiian using the alias Smoothie who mints Litecoins using a similar process to Mike. Physical coins are minted and then tamperproof holographic stickers containing digital addresses are attached to the coins, also similar to Mike, Smoothie doesn’t collect payment in USD or AUD, instead he accepts Litecoins. Smoothie is very interesting in that he will only provide coins that he has already minted. He doesn’t believe in taking orders and then ordering coins. He is a true collector and enthusiast himself and takes the risk that if no-body else wanted to collect his physical coins, that he would keep them all himself; however, given that his small coin runs disappear extremely quickly I don’t believe it is too much of a risk on his part. After an email conversation I placed an order and on Tuesday my physical coins came in. They look great and the boxing is extremely professional. Attached are some of the pictures.

New Book Chapter Preprint

posted Jan 28, 2014, 1:24 PM by Evan Morrison   [ updated Jan 28, 2014, 1:24 PM ]

Happy new year everyone,

It's been a while since I posted an update to the blog, though this one is pretty awesome. Throughout the Decemeber period I was putting the final touches on a new book chapter for Demand-Derive Webservices: Theory, Technologies and Applications edited by the esteemed Dr Zhaohao Sun. The chapter provides a theory for dynamic service composition.

I've attached a preprint to this post, which is not the final version that will be printed in the book and it lacks some of the content of the book chapter version. Most of the parts of the framework discussed in the chapter have been tested on TextSeer. Happy reading.

Evan

November 07th, 2013 - TextSeer updates

posted Nov 6, 2013, 8:14 PM by Evan Morrison

I've just added in a full list of updates that have been rolled into TextSeer over the last year. It's funny because the code base has basically completely been re-written over the last 6-8months.



https://sites.google.com/a/fnord.be/flight-of-the-fnord/textseer/development/2013-11-07-additionsoverthelastyear

November 04th, 2013 - Making a default logic reasoner

posted Nov 6, 2013, 8:01 PM by Evan Morrison

I just wrote a new blog entry about adding a Default Logic Reasoner to TextSeer on my professional webpage - http://www.evanmorrison.com.au/1/post/2013/11/november-04th-2013.html


Running a process environment for teaching BPM

posted Sep 19, 2013, 8:58 PM by Evan Morrison   [ updated Jan 14, 2014, 7:22 PM ]

View what the process lab at the University of Wollongong Looks like Video

In 2010, frustrated with my inability to use industry based process models (due to NDA’s and other privacy concerns) that I was creating during contract engagements I set out to solve my problem in a unique way. This involved speaking to my PhD supervisor Professor Aditya Ghose, who just so happens to teach budding students the art of requirements gathering and process mapping. The idea that I had come up with was that I would create a learning platform for students. I did this by deploying a multi-tenant Oryx BPMN environment. In return for the server configuration,  I was given the ability to direct the assignments specification slightly (I could set broad industry space for the mapping exercises).

After making a few calls and procuring a server I set about configuring an environment that would allow students to do web based process modeling. The result was the “SOA BPMN Toolkit” which is the still used as the classroom environment for all SOA students at the University of Wollongong today. There have been improvements over time and also a few setbacks that have caused hiccups. By and large the system remains relatively beneficial for conscientious students.

What have I learned through this process?

Students do not follow instructions. As part of my deal for a server,  the server was to be inaccessible from outside of the university network. This was a relatively easy request for me to follow as it meant that I did not have to ask to have the port unblocked. However, it did mean that students could not easily log into the server from their homes. It was possible with a little SSH tunnelling (which any computer science student should be able to understand and do with relative ease). To assist students, I provided documentation for logging into the server from offsite.  Surprisingly, year after year there is always at least one student who emails me in the final week before assignments are due complaining that they cannot access the server.

Roughly 99% of students do not do any level of proofing or take steps to learn how to do things for themselves. In the last three years,  I have only had one email from a student who had pushed the system beyond its capabilities. That student wished to understand each of the possible BPMN elements and was using an element that was available in the offline tool provided to all students (in case the server was inaccessible). When he tried to create his process in the online platform, he found that the tool lacked the model element that he wanted to use. He then emailed me and asked if I could help. For his specific request, I dedicated roughly an afternoon verifying and then hand converting his model (from BPMN2 xml into Signavio JSON ecoddings). Interestingly, that student was the only student in his class to go on to begin a research career and is currently enrolled in a PhD program.

Students love excuses. One of the saddest parts of the exercise always occurs in the final week before submission. When roughly 70% of the class wakes up and realises that they have an assignment to submit. Typically this means that all of a sudden the server is overrun and the server crashes under the weight of the students logging in at the last minute. We run a virtual server and do make as many efforts as possible maintain the server online, but without a large budget the server still crashes from time to time. The result of which is an excuse for students completing at the last minute. Instead of downloading the offline tool that is provided (see learning point number 1), the students throw up their hands in defeat and demand extensions for their work, claiming that they are under resourced and that it is absolutely impossible to work in such an environment. Typically, the students are given extensions which they use as much as possible and generally submitting 2-3 weeks late. On reviewing late assignments I often shake my head and wonder what jobs these students will end up taking on. Many assignments are poorly completed tasks with incorrect labelling, construction and non-logical flow – basically a demonstration of everything that they are taught not to do in class. Ultimately these students demonstrate that given 2 months they are still unable to construct even the basic documentation that a typical BA does on a daily basis.

So are we doomed?

No, not really. On average 30% of students in the class ‘get it’ and take the necessary steps to understand the subject. It does mean that we need a way to accelerate the way that students learn industry relative process concepts. My original aim when I started the process laboratory three years ago was to get students to develop processes that were of sufficient quality to use in experimental settings and I think that the way to move forward is to focus on industry specific outcomes for assignment specifications. The subject itself is not supposed to teach students just how to model processes, and is mainly supposed to introduce general concepts using process modeling as a litmus test for the student’s ability to understand a specific aspect of SOA. Moving forward my plan is to roll out EduBPM – a platform for learning process management from an industry perspective.

1-10 of 45